W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2004

Re: Validity Re: [wcag2] Layout tables

From: Maurizio Boscarol <maurizio@usabile.it>
Date: Mon, 19 Apr 2004 18:22:39 +0200
Message-ID: <4083FCCF.6010105@usabile.it>
To: Charles McCathieNevile <charles@w3.org>
Cc: Jens Meiert <jens.meiert@erde3.com>, w3c-wai-gl@w3.org

I could answer Chaals with Gregory's statement "imagining has no place 
in our deliberations as a working group", but I won't do it! :)

I hope anyway my emendation on my last mail has already addressed many 
of the Chaals points. Obviously I largerly agree with him, and I don't 
think validation is not necessary, nor semantic (all Lisa's friends 
included... ;-) ). The use of a simple table for layout, without 
deprecated elements and attributes, preserve both validation and 
semantic of content. The error is in the use of the table, tr, td 
elements in an improper (semantically speacking...) way. But you can 
still validate, navigate through skipping links, levels of headings, and 
so on. At the moment, this way to use table-based layout should be 
indicated as a must (rather than should, or rather than being generic) 
for those who can't or don't want to use css-pure approach, because at 
least this method isn't proved to cause harms - where multiple nested 
table with deprecated elements are to be avoided at all, because they 
can cause very big troubles in navigating and undestranding the page.

We should consider that in some future, any *transitional* tecniques may 
become harmful, or maybe not. But it can always happen, even with better 
standard compliant tecniques, if a major user agent come out with a new 
impredictable bug... God of parsers save us!

Even at a theoretical point of view, anyway, I can't see possible 
problems in navigating into a simple table with good sematic markup and 
jump link inside. But I could lack of imagination... ;-)

I'm talking about known facts, and proposing simply a smarter "plan B" 
approach that can avoid accessibility to be rejected in complex projects 
for bad arguments like "I need/love/am-addicted-to table, table aren't 
accessible, so I won't care about accessibility". I think this is simply 
not true, and we can do our best to indicate that smart uses of layout 
tables for exceptional (and usually large, and important...) projects 
are better than no accessibility at all. Put it in priority 1, put it 
everywhere you like, but don't put it away! :)

"But pandering to broken tools tends to be at the expense of longer-term 
solutions, and I think it is very important to look for alternative 
repair strategies that don't break validation and semantics for others..."

Yes, maybe. In this case I don't think that to have some table-based 
layout in a transitional approach can be at expenses of longer-term 
solutions, because if a site is template driven, at the proper moment 
you can change the template. If a site is static, a well planned, unique 
simple table can be replaced in every page with a simple ad-hoc program 
in minutes. This is not true for complex nested tables, even different 
in every page, with poor semantic inside. To parse and rebuild such a 
site will be a mess. But if we get off people from accessibility due to 
table/table-less myths, we have no chance to see simpler, lighter and 
better built table-based sites.

To stand in the new topic of this spin-off thread, anyway, I have 
absolutely nothing against validation! :)



Charles McCathieNevile wrote:

>Where you might be wrong is if you are prepared to do something that works
>fine for one group of people now, at the expense of others whose problems you
>don't recognise, or at the expense of developing better tools in the future
>that serve people better.
>As an example, consider the idea of navigating through several layers of a
>complex HTML site to a page by skipping through the link text or by skipping
>through the heading structure of each page.
>WCAG contains checkpoints designed to assist both. For a long time people
>were unable to navigate by headings in the most popular tools (Internet
>Explorer, Netscape), although in others such as Opera it was available many
>years ago, providing a useful functionality tyo keyboard users. Instead,
>people would navigate by moving through lists of links.
>This practical reality meant that things like an extra link saying "skip to
>something" would be added at various places in a page. Watching blind JAWS
>users give demos in teh late 90's it was horribly obvious that for them it
>was very difficult to get an overview of the page content because they had no
>mechanism for understanding the heading structure. Yet the absence of this in
>tools was long offered as a reason why it didn't matter if the validation of
>headings made sense, whereas getting link text perfect was considered
>important. (Now JAWS/IE has caught up with other systems and can navigate by
>headings, although it is still somewhat painful in Explorer if you don't have
>JAWS - contrary to what I asserted a while ago the javascript-based
>additional functions tend to have poor structure navigation options.)
>It seems to me that building reliable systems is often the best way to serve
>the most people. And reliable systems tend to require that validation is more
>important overall than a particular presentation aspect.
>As Joe Clark has pointed out, you should be able to build valid, semantically
>correct content and expect users to have tools that implement standards
>(whether they are on a phone, a new shiny machine, a library or school
>computer, using an augmentative systemm to provide a signing avatar, extra
>glossary, or similar comprehension aid...). Reality turns out not to be quite
>like that - people do have poor tools, and cannot always replace them with
>something that works better. But pandering to broken tools tends to be at the
>expense of longer-term solutions, and I think it is very important to look
>for alternative repair strategies that don't break validation and semantics
>for others...
>On Mon, 19 Apr 2004, Maurizio Boscarol wrote:
>>The same difference between serving a validator rather than serving
>>people. Ok, it's a suggestive declaration, I know... ;-) and I have
>>nothing against validators, parsers and such algorithmical tools. But I
>>still prefere people, that's all. If I can't serve both, then I prefer
>>to serve people.
>>Where am I wrong?... :)
>Charles McCathieNevile  http://www.w3.org/People/Charles  tel: +61 409 134 136
>SWAD-E http://www.w3.org/2001/sw/Europe         fax(france): +33 4 92 38 78 22
> Post:   21 Mitchell street, FOOTSCRAY Vic 3011, Australia    or
> W3C, 2004 Route des Lucioles, 06902 Sophia Antipolis Cedex, France
Received on Monday, 19 April 2004 12:22:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:49 UTC